database-cloud-optimization-cost-optimize

Warn

Audited by Gen Agent Trust Hub on Feb 27, 2026

Risk Level: MEDIUMCOMMAND_EXECUTIONEXTERNAL_DOWNLOADSPROMPT_INJECTION
Full Analysis
  • [COMMAND_EXECUTION]: The implementation playbook includes Python code that performs high-privilege operations on AWS infrastructure, such as 'stop_instances' and 'modify_instance_attribute' using the boto3 library (found in resources/implementation-playbook.md). \n- [EXTERNAL_DOWNLOADS]: The skill generates dashboard templates that fetch data visualization scripts from well-known services including jsDelivr and D3js. \n- [PROMPT_INJECTION]: An indirect prompt injection surface exists in the AutoTagger class within 'resources/implementation-playbook.md'. \n
  • Ingestion points: Untrusted CloudTrail event data is ingested via 'event['detail']' in the 'auto_tag_resources' method. \n
  • Boundary markers: No delimiters or safety instructions are present to prevent the model from obeying instructions embedded in the metadata. \n
  • Capability inventory: The skill has the ability to tag any accessible AWS resource with arbitrary string values. \n
  • Sanitization: No validation or escaping is performed on 'userName' or 'sourceIPAddress' before they are interpolated into resource tags, which could allow for metadata poisoning or schema confusion.
Audit Metadata
Risk Level
MEDIUM
Analyzed
Feb 27, 2026, 08:58 AM